EN ▾
SNAPE REP RT Front page →
ConfirmedAI safety03 Nov, 20:10 UTC

Apple's annual filing warns AI agents may take actions nobody authorised and that it may not control them

File photo of an Apple executive speaking on stage at an Apple event.
Getty Images via CNBC

The 10-K risk factors say agents could alter code, send messages or complete transactions without instruction, and that Apple's ability to oversee third-party agents may be limited.

Read at CNBC Share

Why it matters

Risk factors in a 10-K are the legally required list of things that could hurt a company, so the language is a formal admission rather than marketing. It lands on the same day Apple said it would open Siri's AI to all developers, which will let more outside agents act inside its platform. Apple says agents may reach systems they are not authorised to access and that its oversight of third-party agents may be limited. Watch whether other platform owners add similar warnings.

Key facts

  • Warning is in Apple's annual 10-K risk factors
  • Agents 'may behave in unintended or unpredictable ways'
  • Listed risks: altering code, sending communications, completing transactions
  • Oversight of third-party agents 'may be limited'
  • Same day Apple said it will open Siri's AI to all developers

Sources

More on AI safety
More from Snape ReportFront page →